Merge pull request #14 from p2p-ld/fix-hash-printing

print hashes as strings always

Author: sneakers-the-rat

.github/workflows/benchmark.yml

@@ -19,7 +19,7 @@ jobs:
           python-version: ${{ matrix.python-version }}
       - name: Install deps
         run: |
-          python -m pip install -e .[tests]
+          python -m pip install -e .[benchmark]
       - name: Run benchmarks
         run: |
           python -m pytest tests/bench.py --codspeed -vv

docs/changelog.md

@@ -1,5 +1,9 @@
 # Changelog
 
+## Upcoming
+
+- Fix serialization of hashes in `print` mode, which now prints the hex-encoded version everywhere
+
 ## v0.3.*
 
 ### v0.3.3 - 2025-09-30

pdm.lock

@@ -50,10 +50,13 @@ tests = [
     "torrent-models[cli,libtorrent]",
     "pytest>=8.3.5",
     "pytest-cov>=6.0.0",
-    "pytest-codspeed>=3.2.0",
     "torf>=4.3.0",
     "pytest-profiling>=1.8.1",
 ]
+benchmark = [
+    "torrent-models[tests]",
+    "pytest-codspeed>=3.2.0",
+]
 dev = [
     "torrent-models[tests,mypy]",
     "ruff>=0.11.2",
@@ -185,7 +188,7 @@ line-length = 100
 mypy_path = "$MYPY_CONFIG_FILE_DIR/src,$MYPY_CONFIG_FILE_DIR/stubs"
 packages = ["torrent_models"]
 warn_redundant_casts = true
-warn_unused_ignores = true
+warn_unused_ignores = false
 show_error_context = true
 show_column_numbers = true
 show_error_code_links = true

pyproject.toml

@@ -6,7 +6,7 @@
 from urllib.parse import quote
 
 from annotated_types import Ge, Len
-from pydantic import AfterValidator, AnyUrl, BaseModel, Field
+from pydantic import AfterValidator, AnyUrl, BaseModel, Field, PlainSerializer, SerializationInfo
 
 from torrent_models.base import ConfiguredBase
 from torrent_models.types.serdes import ByteStr
@@ -29,8 +29,20 @@ class TorrentVersion(StrEnum):
 """Placeholder in case specific validation is needed for filenames"""
 FilePart: TypeAlias = ByteStr
 """Placeholder in case specific validation is needed for filenames"""
-SHA1Hash = Annotated[bytes, Len(20, 20)]
-SHA256Hash = Annotated[bytes, Len(32, 32)]
+
+
+def _serialize_hash(value: bytes, info: SerializationInfo) -> bytes | str:
+    if info.context and info.context.get("mode") == "print":
+        v = value.hex()
+        if info.context.get("hash_truncate"):
+            v = v[0:8]
+        return v
+    else:
+        return value
+
+
+SHA1Hash = Annotated[bytes, Len(20, 20), PlainSerializer(_serialize_hash)]
+SHA256Hash = Annotated[bytes, Len(32, 32), PlainSerializer(_serialize_hash)]
 
 
 TrackerFields = TypedDict(

src/torrent_models/types/common.py

@@ -9,10 +9,10 @@
 from functools import cached_property
 from math import ceil
 from pathlib import Path
-from typing import TYPE_CHECKING, Annotated, Any, NotRequired, TypeAlias, cast
+from typing import TYPE_CHECKING, Annotated, Any, Callable, NotRequired, TypeAlias, cast
 from typing import Literal as L
 
-from pydantic import AfterValidator, BaseModel, BeforeValidator, PlainSerializer
+from pydantic import AfterValidator, BaseModel, BeforeValidator, WrapSerializer
 from pydantic_core.core_schema import SerializationInfo
 from typing_extensions import TypeAliasType, TypedDict
 
@@ -44,12 +44,11 @@ def _validate_v2_hash(value: bytes | list[bytes]) -> list[bytes]:
     return value
 
 
-def _serialize_v2_hash(value: list[bytes], info: SerializationInfo) -> bytes | str | list[str]:
+def _serialize_v2_hash(
+    value: list[SHA256Hash], handler: Callable, info: SerializationInfo
+) -> bytes | list[str]:
+    ret = handler(value)
     if info.context and info.context.get("mode") == "print":
-        ret = [v.hex() if isinstance(v, bytes) else v for v in value]
-
-        if info.context.get("hash_truncate"):
-            ret = [v[0:8] for v in ret]
         return ret
     else:
         return b"".join(value)
@@ -66,7 +65,7 @@ def _sort_keys(value: dict) -> dict:
 
 
 PieceLayerItem = Annotated[
-    list[SHA256Hash], BeforeValidator(_validate_v2_hash), PlainSerializer(_serialize_v2_hash)
+    list[SHA256Hash], BeforeValidator(_validate_v2_hash), WrapSerializer(_serialize_v2_hash)
 ]
 PieceLayersType = dict[SHA256Hash, PieceLayerItem]
 FileTreeItem = TypedDict("FileTreeItem", {"length": int, "pieces root": NotRequired[SHA256Hash]})

src/torrent_models/types/v2.py

@@ -1,5 +1,18 @@
+import sys
 from pathlib import Path
 
+import pytest
+from _pytest.python import Function
+
 from .fixtures import *
 
 DATA_DIR = Path(__file__).parent / "data"
+
+
+def pytest_collection_modifyitems(config: pytest.Config, items: list[Function]) -> None:
+    # don't run libtorrent tests on windows
+    if sys.platform == "win32":
+        skip_lt = pytest.mark.skip(reason="libtorrent python wheels are bugged in windows!")
+        for item in items:
+            if item.get_closest_marker("libtorrent"):
+                item.add_marker(skip_lt)

tests/conftest.py

@@ -1,6 +1,5 @@
 from pathlib import Path
 
-import libtorrent
 import pytest
 
 from torrent_models import Torrent
@@ -15,6 +14,7 @@ def test_parse_hybrid():
     torrent = Torrent.read("tests/data/qbt_directory_hybrid.torrent")
 
 
+@pytest.mark.libtorrent
 @pytest.mark.parametrize(
     "tfile",
     [pytest.param(tf, id=str(tf.name)) for tf in ALL_TORRENTS],
@@ -23,6 +23,8 @@ def test_infohash(tfile: Path):
     """
     Test that the infohash that we get from a torrent is the same as what libtorrent would compute
     """
+    import libtorrent
+
     lt_torrent = libtorrent.load_torrent_file(str(tfile))
     t = Torrent.read(tfile)
 

tests/test_torrent/test_torrent.py

@@ -3,7 +3,9 @@
 
 import pytest
 
-from torrent_models import KiB, TorrentCreate
+from torrent_models import KiB, Torrent, TorrentCreate
+
+from .conftest import DATA_DIR
 
 
 @pytest.mark.parametrize("version", ["v1", "v2"])
@@ -61,3 +63,19 @@ def test_webseed_url_multifile(version: str, tmp_path: Path):
     # ensure name is prepended to directories
     assert prange.webseed_url("https://example.com/data/") == ws_expected
     assert prange.webseed_url("https://example.com/data") == ws_expected
+
+
+def test_print_hash():
+    """
+    When serialized in `print` mode, hashes should be serialized as strings everywhere
+    """
+    t = Torrent.read(DATA_DIR / "qbt_directory_hybrid.torrent")
+    dumped = t.model_dump(context={"mode": "print"})
+    assert all([isinstance(p, str) for p in dumped["info"]["pieces"]])
+    assert all(
+        [isinstance(f[""]["pieces root"], str) for f in dumped["info"]["file_tree"].values()]
+    )
+    assert all([isinstance(k, str) for k in dumped["piece_layers"]])
+    for k, v in dumped["piece_layers"].items():
+        assert isinstance(k, str)
+        assert all([isinstance(v_hash, str) for v_hash in v])